nonparametric nonlinear control of productivity shocks in estimation of agricultural production function of selected provinces in iran: (case study: wheat production function)

productivity shocks are one of the most important factors that influence the firms ‘ decisions on he optimal consumption of production inputs. unobservable productivity shocks create selection and simultaneity problems causing biased estimates of production function by ordinary least squares. however, to estimate coefficients associated with production function by levinsohn and petrin technique, coefficients associated with intermediate inputs are used as a proxy variable. it is therefore argued that when firms face productivity shocks, intermediate inputs will be moderated by firms. in this approach, because of the nonlinear and nonparametric controlls on productivity shocks, some unbiased estimates could be achieved. to examine this hypothesis, a data set from 11 provinces of iran during the years between 2000 and 2008 was collected. for the collected data set, a production function of wheat was estimated using the three techniques of ols, random effects and levinsohn and petrin method. based on the results, ordinary least square had an upwardly biased estimate of water and fertilizer, while the ols showed downwardly biased estimates of capital. therefore to control the bias due to simultaneity and selection problems, nonparametric methods should be used to model unobservable productivity shocks. in addition result shows that productivity shocks in wheat farms of the selected provinces, are moderated by adjusting the input demand, and thus, there is a significant correlation between input adjustment and productivity shocks.